Researching the politics of development



The art of delivery

Dan Hymowitz
16 February 2016 
Dan Hynmovitz, Head of Insight and Learning and the Tony Blair Africa Government Initiative, share his insights on delivery
Top three messages for governments and international partners

  • Too much science, not enough art: There is too much focus on the technical aspects of delivery, at the expense of political authority and incentives.
  • Start with the context: Begin by thinking about four building blocks of delivery from the centre and how they can be harnessed to drive results across the system.
  • Framing the art of delivery: There’s no best practice for focusing political authority or creating incentives; we offer options based on what we’ve seen work in different places.

In 1941, the United States (US) was ramping up military production to prepare for World War II. At the start, the process was hampered by waste and inefficiency. The Truman Committee – led by then-Senator Harry Truman – was set up to fix this. Truman famously drove 10,000 miles around the US in his Dodge to different military installations to personally check up on things; a hugely successful investigative effort that saved billions of dollars and improved efficiency. Truman might remind us that ‘delivery’ – and by that we mean approaches used by the centre of government to improve policy implementation – isn’t a 21st century invention.
But in international development today there is a growing focus on delivery as part of the response to the fact that many developing countries are stymied by poor governance, from World Bank (WB) President Jim Kim’s call to action on the ‘science of delivery’, to delivery units sprouting up around the world (there are at least 15 at the national and subnational level), to an expanding literature on the topic. For too long, national political systems and the international development industry have focused on policy design at the expense of implementation, but writing the policy is often the easy bit. Poverty reduction doesn’t happen when you press print on a PowerPoint.
The hard part of governing is taking grand visions and plans and turning them into roads, electricity and jobs. The problem is that as delivery has been commodified, it’s in danger of being seen as a silver bullet, and important parts of the original idea are getting lost. Today, the focus is increasingly on the more technical and structural aspects of supporting implementation, such as performance monitoring and setting up delivery units. These things are important, but in AGI’s experience working in 10 African countries over the last seven years, when you ignore the more political facets of delivery, such as the nature of power in any given system and the harnessing of incentives, you won’t get the results you want. There is a risk of using too much science and not enough art in the work of making delivery happen.
Four building blocks of delivery
Delivery units have been phenomenally successful in achieving policy outcomes across the world, from the UK to Malaysia to Australia, and there are a number of exciting innovations underway in Pakistan and Tanzania, as well as, we think, in our own work. But the danger this creates is in thinking it is about the delivery unit itself in all circumstances, as opposed to the principles it represents. We’ve seen delivery units work well in some places but not in others. Therefore, as the WB has pointed out, you shouldn’t default to just setting up a unit regardless of context. In our experience, other mechanisms like stocktakes and performance contracts can sometimes have more impact. To figure out what to do when, governments and international partners should begin by diagnosing how well the existing system is doing across what we see as four building blocks that matter most for delivery from the centre. Two of these elements are part of the ‘science’ and can be fulfilled by a unit or other mechanism, but the other two are more of an ‘art’. The science… and the art…
A way of monitoring performance:
There needs to be a way for the centre to know whether delivery is on- or off-track. This requires clear assignment of responsibility, a mechanism for regularly monitoring progress and the ability to measure.
Political authority:
You need to be able to connect political authority to the delivery process, and this requires an understanding of how power works in a given political system, which is often both complex and nuanced.
A system to respond when things aren’t going to plan:
If delivery is off-track, you need to do something about it. This means identifying the right challenges and then responding to them. A delivery unit, or similar, can be effective in identifying barriers and solutions.
Policy delivery often requires actions by a range of actors across a number of organisations, and for this to happen there need to be positive and negative incentives – often both formal and informal – to make sure government leaders and officials are focused and motivated.
The art of delivery: political authority and incentives
Governments and international partners today risk repeating mistakes of the past by focusing too much on the scientific parts of delivery. This is because it is more comfortable territory for technical assistance and donor programmes. But monitoring and problem solving won’t work without the application of political authority and aligned incentives. In one country we’ve worked with, the centre, working with international partners, has introduced complex tracking tools that ministries are indeed filling out each week.
However, that doesn’t mean the system works. In one of the main ministries, for example, we’ve observed that in senior meetings, the minister doesn’t pay attention to the data generated through this process. Weekly reports which show that most of the ministry’s initiatives are off-track are therefore getting ignored. This risks becoming what Lant Pritchett, Michael Woolcock and Matt Andrews have called “isomorphic mimicry”: it’s a delivery process on paper and even in terms of activities; but without the correct authority and incentives, it’s a system that doesn’t improve actual implementation and decision-making. Nobody is less poor as a result.
The filled out tracker is just as useless as the countless white papers gathering dust on the minister’s shelf. The important question for governments and international partners who want better results is how to harness political authority and incentives. These two are related: effectively connecting political authority to the delivery process is one of the best ways to create informal incentives. But because doing this well depends on context, it’s not as simple as, for example, always using performance contracts. We certainly don’t have all the answers, but below we talk about some of the things that have been effective in the countries we work with, and some that haven’t.
Political authority
The question here is how to focus political authority on the delivery process. When this is done well, you can create incentives to deliver and unblock problems. But just because it should happen doesn’t mean it always does in complex political systems, and when this is missing, the system won’t work. In one country AGI has worked in, delivery unit staff didn’t feel they had the authority to challenge ministers because the unit didn’t have the full backing of the president. The result was a body that went through the motions of ‘monitoring’ progress without any real teeth. Before deciding how to link political authority to the delivery process, it’s essential to first understand where power resides and how it is exercised, which often relies on informal networks as much as formal hierarchies. In some governments, the answer is simple: the president. But authority isn’t always where the organisational chart says it is. In many countries, political authority is diffused and political parties, actors outside of the official system or groups of key ministers may be the true power brokers. Understanding this should be the starting point for the design of any system, and because power is dynamic, not static, any system will need to be able to adapt over time. We’ve seen four approaches to using central political authority to drive implementation.
Involving political leaders directly in the delivery process
This is in some ways the most literal, and there are different ways to do it. As with other approaches we talk about in this paper, the appropriate method to use will vary depending on country and situation. One core technique is stocktakes. These are regular, standardised progress updates for the leader, and often take the form of meetings between the president and the minister in charge of a priority issue. The stocktake process allows the president to keep pressure on the minister and also to intervene to assist when there are holdups or blockages. An approach that’s potentially complementary to stocktakes is performance contracts. These are contracts signed by ministers and other senior officials that commit them to specific delivery targets.
One version we’ve seen work well is Imihigo in Rwanda. This is an adapted version of a cultural tradition in the country that leads to a contract with the president committing the government official to specific deliverables. Officials in Rwanda credit Imihigo with successes in health, such as a huge increase in health insurance coverage, which is now above 90 percent nationwide, road construction and agricultural production. But the more general point is to find ways for the leader to step up their engagement with the issues they care about. For example, in Sierra Leone in 2010, President Koroma was struggling to find consistent space in his diary to engage with his priorities. The president set up a weekly ‘delivery day’ – a particular day each week where his calendar was blocked off for just those issues.
In practice, events often intervened and made this difficult to stick to, but it helped him to stay consistently involved, and there was value to the signal of having the delivery day in place. Nigeria’s current president and his predecessor have both tried different versions of this. President Jonathan inserted himself as chair of a national task force on power, which he led each week to make sure he was personally engaged on his top issue. President Buhari, meanwhile, has made himself oil minister, giving him more control, but also signalling the importance of a critical area for Nigeria.
Using delegated political authority
Political authority isn’t synonymous with the president. Another way to go is to have someone closely connected to the president oversee day-to-day performance management. This person may be a chief of staff or the head of a dedicated delivery unit. In Nigeria, however, President Jonathan made Finance Minister Dr Ngozi Okonjo-Iweala the Coordinating Minister for the Economy (CME), and she led the Economic Management Implementation Team (EMIT), which steered progress on the delivery of the president’s Transformation Agenda.
Harnessing collective authority
When political authority is diffused, rather than in one person, it’s even trickier to design a delivery mechanism. This is where the art of delivery comes in. One approach is for a senior figure to personally engage. In one country we’ve worked in, tensions between the finance minister and the chief of staff – both of whom were rising stars in the governing party and essential to delivery and cross-government coordination – were holding up implementation of projects.
Recognising this challenge, a new mechanism was set up. A meeting was held every two weeks between the president, chief of staff, Ministry of Finance and a small group of advisors to review progress on 10 priority projects. This helped keep momentum on the president’s top initiatives by building key people into the process. Another option is to align people behind a common agenda or project. In Liberia in early 2015, progress in the fight against Ebola was being derailed because of ongoing disagreements between three centres of power: the head of Monrovia’s Ebola task force, the head of contact tracing, and important international partners. This group began to work together more effectively only after a new compromise strategy was designed that each of these groups agreed with. Within two months Liberia would be declared Ebola free, in part because of the genuine collaboration behind this strategy.
Helping create outside political pressure
Symbolic measures – such as a speech or public event – to emphasise to those doing the implementing and to the public that the president is committed to delivery of the priority can be another powerful approach. For example, President Koroma made a speech in November 2009 announcing that a free maternal and child health programme would launch in April 2010. This bold statement both signalled President Koroma’s seriousness to the rest of government and created pressure on Koroma to live up to his promise. The free healthcare initiative would eventually launch on time; one of the few countries in which this happened.
There need to be positive and negative incentives to make sure government leaders and officials are focused and motivated, and it is political leaders who often create them. The trickiest part is setting up incentives that both encourage government officials to deliver and to participate in the performance management process in a constructive way. When many people talk about incentives, they focus on more technical things, like pay and promotions. These are obviously important, and equally one cannot ignore the strong incentives created by corruption, especially in low pay environments. But there are also more subtle approaches, like presidents rewarding high performing ministers with more face time or public praise.
In the UK, Michael Barber used to talk about “Prime Ministerial stardust” or the buzz a civil servant would get simply from walking through the famous black door to Number 10. In our experience, getting these soft and informal incentives right is more art than science. It is context dependent and might require mixing and matching different tools, but can be highly effective in supporting change. One way to frame the options is to think about a two-by-two matrix. On the vertical axis, there’s the choice between positive incentives (carrots) and negative incentives (sticks). On the horizontal axis is the option of mobilising incentives from within government, as opposed to using incentives from outside of government.
If performance management were a science, Truman’s driving around the US to personally inspect production camps looks absurdly inefficient. Why not get data sent in and analyse it back in Washington? Or use a team of inspectors to fan out across the country? But there were reasons why Truman’s way was better. The element of surprise when he showed up unannounced at the production camps meant that he was getting an accurate picture of what was happening. In addition, the symbolism of his cross-country drive and the credibility he gained from this allowed him to assertively lead the committee hearings and take on powerful actors in the US defence industry. In other words, Truman wouldn’t have been successful without using a bit of art too. Delivery units and related systems can be a great way of driving change in complex government systems. They provide focus, energy, and the data to inform decision-making. But no unit or system is ever a silver bullet and will fail without an alignment of authority and incentives. This needs to be borne in mind in donor programming and government planning if we are not to have too much science and not enough art.
This post was  originally featured on